Re: [PATCH] x86emul: don't unconditionally clear segment bases upon null selector loads

[Andrew Cooper <andrew.cooper3@citrix.com>]

On 20/12/2016 08:18, Jan Beulich wrote:
> AMD explicitly documents that namely FS and GS don't have their bases
> cleared in that case, and I see no reason why guests may not rely on
> that behavior. To facilitate this a new input field (the CPU vendor) is
> being added.
>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

This looks better overall.

Longterm I think it would be better to pass the full cpuid policy in to
the emulator. This removes the need to use the cpuid() hook for both
emulation and instruction related purposes, which we have seen gets
complicated with CPUID Faulting handling.  Looking further than that,
passing the full MSR banks would simplify that side of things as well.

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>, with one minor
correction

> --- a/xen/arch/x86/hvm/emulate.c
> +++ b/xen/arch/x86/hvm/emulate.c
> @@ -1897,6 +1897,7 @@ void hvm_emulate_init_once(
>      hvmemul_get_seg_reg(x86_seg_ss, hvmemul_ctxt);
>  
>      hvmemul_ctxt->ctxt.regs = regs;
> +    hvmemul_ctxt->ctxt.vendor = current->domain->arch.x86_vendor;

curr is available here.

~Andrew

>      hvmemul_ctxt->ctxt.force_writeback = true;
>  
>      if ( cpu_has_vmx )
>


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel